- From: Kevin Regan <kevinr@valicert.com>
- Date: Wed, 12 Jul 2000 13:27:13 -0700 (PDT)
- To: w3c-ietf-xmldsig@w3.org
This question is in relation to the KeyInfo structure. What I've gotten from the spec and from various emails is that at least one KeyValue corresponding to the actual authentication key must be included with the KeyInfo. However, I'm wondering if it would make sense to make this more explicit. Currently, the KeyInfo element looks like this: <element name='KeyInfo'> <complexType content='elementOnly'> <choice minOccurs='1' maxOccurs='unbounded'> <any namespace='##other' minOccurs='1' maxOccurs='unbounded'/> <element name='KeyName' type='string'/> <element ref='ds:KeyValue'/> <element ref='ds:RetrievalMethod'/> <element ref='ds:X509Data'/> <element ref='ds:PGPData'/> <element ref='ds:SPKIData'/> <element name='MgmtData' type='string'/> </choice> <attribute name='Id' type='ID' use='optional'/> </complexType> </element> Instead, maybe you could use something like: <element name='KeyInfo'> <complexType content='elementOnly'> <element ref='ds:KeyValue'/> <choice minOccurs='1' maxOccurs='unbounded'> <any namespace='##other' minOccurs='1' maxOccurs='unbounded'/> <element name='KeyName' type='string'/> <element ref='ds:KeyValue'/> <element ref='ds:RetrievalMethod'/> <element ref='ds:X509Data'/> <element ref='ds:PGPData'/> <element ref='ds:SPKIData'/> <element name='MgmtData' type='string'/> </choice> <attribute name='Id' type='ID' use='optional'/> </complexType> </element> In this definition, there would always be a KeyValue for the authentication key. Any additional information, such as certificate chains, could be included in addition to this. However, the KeyValue would always be available and unambiguously identify the authentication key. My take on the majority of the Signature subtree is that any application can look at a Signature generated by a different application and verify it, as long as the key is available. Where this breaks down is in actually obtaining the key -- the structure of SignedInfo is much more strictly specified than that of KeyInfo. Again, any effort made to standardize this would be appreciated. --Kevin
Received on Wednesday, 12 July 2000 16:27:10 UTC